Download A reassessment of the timing of early Archaean crustal

A reassessment of the timing of early Archaean crustal
evolution in West Greenland
Stephen Moorbath and Balz Samuel Kamber
In last year’s Review of Greenland activities, Kalsbeek
(1997) divided the recent history of geochronology into
three successive periods:
1. single-sample K-Ar and Rb-Sr mineral or whole-rock
age determinations;
2. Rb-Sr and Pb/Pb whole-rock isochrons and multigrain
zircon U-Pb isotope data;
3. the present, where ‘single’ zircon U-Pb data are predominantly used.
To these three, we would propose adding a fourth,
namely a combination of all three, in order to achieve
the maximum age information within complex terrains.
For an early Precambrian terrain like that of West
Greenland, we consider that the combined use of at least
the last two approaches is essential (to which should
be added the Sm-Nd method). In recent years, study
of the geochronological evolution of the Godthåbsfjord
and Isua regions has been dominated by rapid and precise ion-probe U-Pb dating of complex-structured zircons, and it has become fashionable to regard the wide
range of zircon dates, and particularly the oldest, as giving the age of rock formation. Dates obtained from
whole-rock Rb-Sr, Sm-Nd and Pb/Pb regressions have
been regarded as too imprecise for adequate age resolution, whilst constraints on crustal evolution imposed
by initial Sr, Nd and Pb isotope ratios have been summarily dismissed or totally ignored. We consider that
this sole dependence on ion-probe dating of zircon can
lead (as, indeed, in the early Archaean of West Greenland) to a potential misinterpretation of the timing of
crustal evolution, especially in those cases where little
or no information regarding the relationship between
measured date and internal grain structure is available.
Figure 1 shows the localities mentioned in the text.
100 km
0
25 km
ISUA
N
66 °N
Nordlandet
64°
God
thåb
sfjo
TRACT
CONTAINING
AMITSOQ GNEISSES
rd
Archaean
block
Angisorsuaq
Akilia
N
ar
ss
aq
Ame
ralik
48 °W
Nuuk
Innersuartuut
52°
Buksefjorden
64°
88
50°
Fig. 1. Sketch-map of the area around
Nuuk, West Greenland with localities
mentioned in the text.
Geology of Greenland Survey Bulletin 180, 88–93 (1998)
© GEUS, 1998
15.0
14.5
3.0 Ga
Pb/
204
Pb
14.0
3.5 Ga
13.5
3.6 Ga
207
Fig. 2. Common Pb diagram with mantle
evolution line after Kramers & Tolstikhin
(1997). Black open squares represent 83
published whole-rock data points which
regress to 3654 ± 73 Ma (MSWD = 17.6)
and intersect the mantle evolution line at
3.66 Ga. Red full squares correspond to
whole rock, HF-leached feldspar, and
feldspar leachate analyses of four
Amîtsoq gneisses claimed to have ages in
the range 3.82–3.87 Ga from U-Pb zircon
dates (for sampling details see text).
Mantle-derived Pb of 3.85 Ga (stippled
line) would be expected to lie on a
quasi-parallel trend to the bulk Amîtsoq
data, but off-set down towards older
model intersection ages. Blue full circles
are leached plagioclase and whole rock
data from gabbroic Akilia Association
enclaves. Some of these enclaves are cut
by gneiss sheets for which ages of
3.85–3.87 Ga have been claimed. Their
Pb isotope composition offers no
evidence for such an old age but is
compatible with a 3.67 Ga Sm-Nd
isochron age (Fig. 3) obtained on similar
samples.
3.7 Ga
13.0
3.8 Ga
3.9 Ga
12.5
4.0 Ga
12.0
10.0
Whole-rock regression ages for the
Amîtsoq gneisses
We have regressed all available whole-rock isotopic
data for the Amîtsoq gneisses (source references available from the authors), to yield concordant Rb-Sr (3660
± 67 Ma), Sm-Nd (3640 ± 120 Ma), and Pb/Pb (3654 ±
73 Ma) ages. These regressions are by no means perfect isochrons, and the scatter of points about the regressions in excess of analytical error is due to: (1)
open-system behaviour for parent or daughter isotopes,
or both, during well-attested late Archaean and midProterozoic metamorphism; (2) a small degree of heterogeneity in initial Sr, Nd and Pb isotope ratios for
different components of the Amîtsoq gneisses; (3) a
combination of these. However, we regard the weighted
mean age of 3655 ± 45 Ma (2 sigma error) from all
three methods as a reliable estimate for the emplacement age of the magmatic precursors of the Amîtsoq
orthogneisses. It is improbable that agreement between
these three methods is simply fortuitous, or the result
of some massive, regional metamorphic or metasomatic
event. Furthermore the initial Sr, Nd and Pb isotopic constraints are also concordant, all strongly indicating a
10.5
2σ
11.0
11.5
12.0
206
Pb/
204
12.5
13.0
13.5
14.0
Pb
mantle-like source, rather than much older, reworked
sialic crust.
All Amîtsoq gneisses studied by the ion-probe U-Pb
technique have yielded at least some zircon dates in
this range, and c. 3.65 Ga is seen as the age of a major
crust-forming event by Nutman et al. (1993, 1996).
However, these workers have also reported many older
zircon dates from which it is concluded that the gneiss
complex had a complicated earlier history, having been
added to, and modified, in several events starting at
c. 3900 Ma and extending down to c. 3600 Ma. Fortunately, it is possible to test these claims independently
by combining information obtained by ion-probe U-Pb
dating and Pb-isotope systematics (the combined
approach 4). Of particular interest is the question of
whether the zircon dates really refer to the true age of
formation of their host rock, or whether they only refer
to the age of the zircon itself. In the latter case, it would
have to be concluded that the zircon is inherited from
an older rock which may no longer be exposed. Zircon
is known to be an extremely hardy, resistant mineral
which can survive sedimentary and magmatic cycles (e.g.
Lee et al. 1997; Mezger & Krogstad 1997).
89
Pb-isotopic constraints for the age of
the Amîtsoq gneisses
The evolution of Pb isotopes in continental crust, oceanic
crust, mantle and meteorites through earth’s history has
been closely studied for over forty years. Part of the primary isotopic growth curve for mantle Pb, which links
the most primitive Pb of iron meteorites with modern
mantle-derived Pb, is shown in Figure 2. The shape of
this curve, as well as the increasing resolution towards
older ages, is due to the very different half-lives of 235U
(703.8 Ma) and 238U (4468 Ma). Here we use the mantle evolution curve of Kramers & Tolstikhin (1997),
which barely differs in the relevant time range from the
well-known primary growth curve of Stacey & Kramers
(1975). It should be noted that the 207Pb/206Pb compositions of 3.85 Ga and 3.65 Ga-old mantle Pb differ by
13%, far outside any analytical uncertainties (typically
c. 0.15%).
All 83 Amîtsoq gneiss common leads so far analysed
by various workers fall within the data envelope shown
in Figure 2. The data points scatter around a regression
line which yields an age of 3654 ± 73 Ma. Most present-day Amîtsoq gneiss leads are extremely unradiogenic (because of the low U/Pb ratio of the gneisses),
and fall between 11.5 and 12.5 on the 206Pb/204Pb scale
(Fig. 2). This is the main reason for the fairly high age
error on the regression line. Of much greater, indeed
crucial, importance is that the Amîtsoq gneiss Pb-isotope regression intersects the mantle evolution curve
precisely at 3.66 Ga. This is, within error, identical to
the 3.65 Ga intercept with the earlier growth curve of
Stacey & Kramers (1975). From this we conclude that
the magmatic precursors of all the analysed Amîtsoq
gneisses were derived from the mantle, or from a geochemically similar source, at 3.65–3.66 Ga, after which
they became part of the continental, granitoid crust.
There is simply no hint of the presence of any Amîtsoq
gneiss which began its existence in the crust as long
ago as c. 3.85 Ga (Fig. 2).
It might be argued that all of the above gneisses
would yield ion-probe U-Pb zircon dates of c. 3.65 Ga.
Unfortunately, few such comparisons are available. We
have therefore included in our common Pb isotopic
studies several Amîtsoq gneisses which yield much
older ion-probe zircon dates as far back as c. 3870 Ma,
for each of which the oldest measured ion-probe date
is interpreted as the true age of rock formation (e.g.
Nutman et al. 1997a). Of particular interest are Amîtsoq
gneisses in and around the island of Akilia, about 25
km south of Nuuk, where ion-probe dates in the range
90
of 3872–3619 Ma have been reported within a small area
(Nutman et al. 1996, fig 2.). On Akilia itself, discordant
sheets of Amîtsoq gneiss with the oldest ion-probe dates
of 3860–3870 Ma cut metasedimentary and meta-igneous
rocks of the so-called Akilia Association, which are thus
regarded as even older (Nutman et al. 1997a). Here an
enclave of highly metamorphosed banded iron formation contains accessory apatite with graphite inclusions
yielding a C-isotope signature regarded as biogenic in
origin (Mojzsis et al. 1996). Consequently, Nutman et
al. (1997a) claim that life existed on earth prior to 3860
Ma, and might therefore have overlapped with a time
when the earth was still being affected by major impacts,
such as probably terminated on the moon at c. 3.80 Ga.
Overlap of truly major impacts with the existence, or
origin, of life is regarded as highly improbable (e.g.
Maher & Stevenson 1988; Sleep et al. 1989). This is further discussed below.
Pb-isotopic compositions have been measured on
whole rocks and feldspars for the following samples with
ion-probe U-Pb zircon dates >>3650 Ma: (1) discordant
sheets of Amîtsoq gneiss on Akilia, as described above,
(2) an Amîtsoq gneiss (GGU 110999) from the island of
Angisorsuaq, 2 km west of Akilia, which has been much
analysed for 25 years, and which was the first such
sample to give an ‘old’ ion-probe date of 3820 Ma
(Kinny 1986), (3) gabbroic Akilia Association enclaves
from Akilia and the nearby (10 km to the south) island
of Innersuartuut, which are older than the discordant
Amîtsoq gneiss sheets with their respective ion-probe
zircon dates of 3865 Ma (Nutman et al. 1997a) and 3784
Ma (Bennett et al. 1993).
Figure 2 shows that the Pb-isotopic compositions of
all these samples fall exactly on the regional Amîtsoq
field which regresses at 3654 ± 73 Ma and, much more
significantly, intersects the mantle evolution curve at 3.65
to 3.66 Ga. We conclude that the magmatic precursors
of even those Amîtsoq gneisses which yield ion-probe
U-Pb zircon dates >> 3650 Ma are the products of a
major mantle-crust differentiation episode at around
3.65 Ga (e.g. Moorbath & Taylor 1981). There is no
sign from the Pb isotopes that crustal development of
any of these Amîtsoq gneisses or gabbroic Akilia enclaves
began as early as the times given by the ion-probe
dates (Fig. 2).
The inescapable corollary from these results is that
zircons significantly older than c. 3.65 Ga are inherited
grains from some older, evolved, regional crust of as
yet unspecified type, which may no longer be exposed.
Our analysis of all published ion-probe data (Kamber
& Moorbath 1998), together with recent ion-probe data
0.5135
0.5130
Nd/
144
Nd
0.5125
143
Fig. 3. Sm-Nd isochron plot for data of
Bennett et al. (1993) on Akilia Association
enclaves from the island of Akilia (one
sample – the highest point), and the
nearby island of Innersuartuut (three
samples). The mean square weighted
deviate (MSWD) of < 1 shows that this is
a statistically perfect isochron. εNd is a
measure of the initial 143Nd/144Nd ratio,
which is of great importance for petrogenetic and geochemical studies as well
as for modelling mantle evolution. The
quoted error on the age is 2 sigma (95%
confidence level).
0.5120
0.5115
Age = 3677 ± 37 Ma
0.5110
MSWD < 1
εNd
0.5105
0.5100
0.08
0.10
0.12
0.14
0.16
147
Sm/
(Nutman et al. 1997b) on detrital zircons from a
metaquartzite in the Isua supracrustal belt, some 150
km north-east of Nuuk, suggests that an event at c. 3.85
Ga is of particular regional importance. Thus, whilst the
discovery of >>3650 Ma-old zircons with the ion-probe
has been of great importance, we consider that they do
not date the time of formation of the rocks which
presently host them. The geochemical nature of the c.
3.85 Ga-old source rocks is difficult to constrain.
Contamination of the younger melts with older material (i.e. the source-rocks of the c. 3.85 Ga-old zircons)
had minimal effects on the Pb-, Nd-, and Sr-isotope
systematics (although some of the scatter around the
regressions might perhaps be explained as stemming
from very minor contamination with country-rock).
Detailed multidisciplinary work on the ancient zircons
themselves will hopefully elucidate the nature of their
source rocks.
Significance for age of earliest life
Our re-interpretation of a rock formation age of c. 3.65
Ga for discordant Amîtsoq gneiss sheets on Akilia provides a new, less spectacular, minimum age for those
Akilia Association enclaves which bear C-isotope evidence for possibly biogenic processes (Mojzsis et al.
1996). The fact that Akilia Association enclaves from
Akilia and Innersuartuut fall on an indistinguishable
Pb-isotopic trend from the discordant (and other)
Amîtsoq gneisses means that the enclaves cannot be
more than a few tens of millions of years older than
the gneisses (provided they were derived from a man-
144
0.18
= +2.6
0.20
0.22
0.24
Nd
tle-like source, which seems likely given their gabbroic
composition and association with ultramafic rocks).
There is some published, independent evidence for
this, which we now discuss briefly.
Bennett et al. (1993) reported Sm-Nd data for a suite
of gabbroic enclaves of the Akilia Association, including Akilia and Innersuartuut. Using minimum age constraints obtained from ion-probe U-Pb zircon data in the
range of 3872 to 3784 Ma from discordant and enclosing Amîtsoq gneisses (see above), they calculated initial Nd isotope ratios for the Akilia gabbros and, together
with analogous comparative data for the Amîtsoq
gneisses, arrived at a model of major Nd-isotope heterogeneity in the earth’s mantle in early Archaean times.
This approach, which has been strongly criticised by
Moorbath et al. (1997), assumes that every analysed
rock remained a closed system to Sm or Nd diffusion
since the time given by the U-Pb zircon date. But were
the bulk rocks already in existence at the time given
by the U-Pb zircon dates? If one plots the Sm-Nd data
of Bennett et al. (1993) for five separate localities (seven
data points) of Akilia Association gabbroic enclaves,
they yield an isochron age of 3675 ± 48 Ma. Plotting
only the data from Akilia (one sample) and Innersuartuut
(3 samples) yields a perfect Sm-Nd isochron (MSWD
< 1) with an age of 3677 ± 37 Ma, as shown in Figure
3. It is probable that this is a close estimate for the age
of not only the gabbroic enclaves on these islands, but
also for the closely associated banded iron formation
lithologies which (on Akilia) contain apatite with graphite
inclusions of probable biogenic origin (Mojzsis et al.
1996). It should be remembered that on Akilia, the
Akilia Association enclaves are cut by a gneissic gran-
91
itoid sheet which yields zircon U-Pb dates up to 3870
Ma, which we regard as inherited zircons from a time
when the present host rocks did not even exist.
The only ion-probe zircon U-Pb date so far measured directly on an Akilia Association rock, namely a
schist from Innersuartuut, was reported by Schiøtte &
Compston (1990). They obtained a complex age pattern, but favoured 3685 ± 8 Ma as representing the original age of this part of the Akilia Association and found
no zircons approaching the value of c. 3865 Ma obtained
by Nutman et al. (1997a) for the discordant gneiss sheets
on Akilia.
Direct age constraints on the Akilia Association
obtained with three independent methods thus yield
a concordant deposition age: (1) Pb/Pb model age constraints on gabbroic samples indicate a mantle extraction age between 3.70 and 3.65 Ga; (2) Sm-Nd analyses
of similar gabbroic samples yield an isochron age of
3677 ± 37 Ma, and (3) a volcanogenic Akilia Association
schist was dated at 3685 ± 8 Ma with the U-Pb ion-probe
method. The combined age of c. 3.67–3.68 Ga is in direct
conflict with the interpretation of U-Pb zircon age spectra of younger, cross-cutting gneiss sheets, which were
believed to be as old as 3.87 Ga (Nutman et al. 1997a).
However, a closer inspection of the age spectra, in
other words the data themselves, reveals that an alternative interpretation is equally plausible. The ion-probe
U-Pb zircon age spectra of all three analysed discordant gneiss sheets can be statistically analysed to yield
between two and four age populations per spectrum
(Nutman et al. 1997a, table 3). No matter which analysis is preferred, prominent, co-existing age populations are always found in the range of 3.81–3.86 Ga
and 3.60–3.65 Ga. Whilst Nutman et al. (1997a) prefer
to view the older population as representing the rock
formation age, the combined geochronological evidence in fact clearly shows that the younger 3.60–3.65
Ga population corresponds to the rock formation age
and that the 3.81–3.86 Ga population was inherited. Our
re-interpretation is not only compatible with the direct
age constraints on Akilia Association rocks but also
with the aforementioned Pb isotope characteristics of
the cross-cutting Amîtsoq gneiss sheets, thereby
demonstrating that the most reliable age constraints
in complex gneiss terrains are obtained by a combination of geochronological techniques, rather than by
application of only one (the most precise) geochronometer.
The revised dates presented here for Akilia Association
enclaves of possible significance for the study of earliest life are nearly 200 Ma younger than the minimum
92
date of c. 3865 Ma proposed by Nutman et al. (1997a).
If our re-interpretation is correct, the question of overlap of earliest life with a lunar-type impact scenario terminating at c. 3.80 Ga, as suggested by Nutman et al.
(1997a), becomes irrelevant.
Space limitations do not allow discussion here of the
age of the Isua greenstone belt. However, we agree
with Nutman et al. (1997b) that deposition of the major
part of the belt probably occurred at c. 3.71 Ga.
This paper summarises a major conflict between current interpretations of the geochronological evolution
of the early Archaean complex of the Godthåbsfjord
region of West Greenland. We trust that future work,
combining approaches (2) and (3) of Kalsbeek (1997),
will resolve this controversy.
Acknowledgements
Moorbath thanks Peter Appel, Chris Fedo, Vic McGregor, Steve
Mojzsis and John Myers for help in the field. Kamber is funded by
a Swiss BBW grant (95.0833). This is a contribution to the Isua
Multidisciplinary Research Project, supported by the Danish Natural
Science Research Council, the Geological Survey of Denmark and
Greenland, the Commission of Scientific Research in Greenland,
and the Minerals Office of the Greenland Government (now Bureau
of Minerals and Petroleum).
References
Bennett, V.C., Nutman, A.P. & McCulloch, M.T. 1993: Nd isotopic
evidence for transient, highly depleted mantle reservoirs in
the early history of the Earth. Earth and Planetary Science
Letters 119, 299–317.
Kalsbeek, F. 1997: Age determinations of Precambrian rocks from
Greenland: past and present. Geology of Greenland Survey
Bulletin 176, 55–59.
Kamber, B.S. & Moorbath, S. 1998: Initial Pb of the Amîtsoq gneiss
revisited: implication for the timing of early Archaean crustal
evolution in West Greenland. Chemical Geology 150, 19–41.
Kinny, P.D. 1986: 3820 Ma zircons from a tonalitic Amîtsoq gneiss
in the Godthåb district of southern West Greenland. Earth and
Planetary Science Letters 79, 337–347.
Kramers, J.D. & Tolstikhin, I.N. 1997: Two terrestrial lead isotope
paradoxes, forward transport modelling, core formation and
the history of the continental crust. Chemical Geology 139,
75–110.
Lee, J.K.W., Williams, I.S. & Ellis, D.J. 1997: Pb, U and Th diffusion in natural zircon. Nature 390, 159–162.
Maher, K.A. & Stevenson, D.J. 1988: Impact frustration of the origin of life. Nature 331, 612–614.
Mezger, K. & Krogstad, E.J. 1997: Interpretation of discordant UPb zircon ages: An evaluation. Journal of Metamorphic Geology
15, 127–140.
Mojzsis, S.J., Arrhenius, G., McKeegan, K.D., Harrison, T.M.,
Nutman, A.P. & Friend, C.R.L. 1996: Evidence for life on Earth
before 3800 million years ago. Nature 384, 55–59.
Moorbath, S. & Taylor, P.N. 1981: Isotopic evidence for continental
growth in the Precambrian. In: Kröner, A. (ed.): Precambrian
plate tectonics, 491–525. Amsterdam: Elsevier Scientific Publishing Company.
Moorbath, S., Whitehouse, M.J. & Kamber, B.S. 1997: Extreme Ndisotope heterogeneity in the early Archaean – fact or fiction?
Case histories from northern Canada and West Greenland.
Chemical Geology 135, 213–231.
Nutman, A.P., Friend, C.R.L., Kinny, P.D. & McGregor, V.R. 1993:
Anatomy of an Early Archean gneiss complex: 3900 to 3600
Ma crustal evolution in southern West Greenland. Geology 21,
415–418.
Nutman, A.P., McGregor, V.R., Friend, C.R.L., Bennett, V.C. &
Kinny, P.D. 1996: The Itsaq Gneiss Complex of southern West
Greenland; the world’s most extensive record of early crustal
evolution (3,900–3,600 Ma). Precambrian Research 78, 1–39.
Nutman, A.P., Mojzsis, S.J. & Friend, C.R.L. 1997a: Recognition of
≥3850 Ma water-lain sediments in West Greenland and their
significance for the early Archaean Earth. Geochimica et
Cosmochimica Acta 61, 2475–2484.
Nutman, A.P., Bennett, V.C., Friend, C.R.L. & Rosing, M.T. 1997b:
~3710 and ≥3790 Ma volcanic sequences in the Isua (Greenland)
supracrustal belt; structural and Nd isotope implications.
Chemical Geology 141, 271–287.
Schiøtte, L. & Compston, W. 1990: U-Pb age pattern for single
zircon from the early Archaean Akilia Association south of
Ameralik fjord, southern West Greenland. Chemical Geology
80, 147–157.
Sleep, N.H., Zahnle, K.J., Kasting, J.F. & Morowitz, H.J. 1989:
Annihilation of ecosystems by large asteroid impacts on the
early Earth. Nature 342, 139–142.
Stacey, J.S. & Kramers, J.D. 1975: Approximation of terrestrial lead
isotope evolution by a two-stage model. Earth and Planetary
Science Letters 26, 207–221.
Authors’ address:
Department of Earth Sciences, Oxford University, Parks Road, Oxford OX1 3PR, UK.
93